Certainly 3Lz+3T should be worth more than 2A+3T -- and it is. Now, should scoring of jump combinations reflect the difficulty of combining each two particular jumps in that order better than they do now. That would be ideal, if there could be a clean way to build in appropriate rewards for the difficulty of each possible combination, independent of what the skater is doing in the other jump passes of that program. Any suggestions?
Sounds like the difficulty is with 2-3 combinations as opposed to 3-2's, then also very fancy combinations e.g. Alina Zagitova's 3Lz-3Lo can only get a GOE bonus of 2.1 marks e.g. the same as a 2A-3T and even a solo 3T. However the proposed +/-5 GOEs and GOE bonuses based on percentages may come to the rescue to some extent.
Re the 2-3's you often get the suggestion on these pages for a 10% bonus for the -3Lo part of a -3Lo combination, and I would be in agreement with that. The -3Lo does seem very difficult to do straightaway, Alina is really the exception and even then she can make mistakes. Re a -3T, do we really want to start adding bonuses for these, they're fairly ubiquitous, and we may have to live with what we've got unless former experienced skaters can say exactly what marks they would give for the various 2-3T combinations (I haven't considered 3 part combos in my thoughts), and if they're more than say 0.5 marks different than what you get now then you might consider a separate table for these, but all the 3 part combos might be a bit too much.
Re fancy combinations it does seem a bit odd that the GOE is only based on the highest base value jump in the combination, and not a sum of the individual GOEs. I can only think this is because 'The Powers That Be' don't want it to be all about combinations which is what you might get if you summed them e.g. Alina's combination could get up to 4.2 points in GOE, and in a 50% +5 GOE world you'd be talking about 5.55 marks. Also the GOE marks for the jumping passes would be much higher than now. At the moment the max is 14.7 marks in 7 jumping passes - 7 lots of 2.1 with two 2-3 combos with or without a -1Lo-, but with 50% possible you'd be talking 22/23 marks with typical base value routines of 44-46 marks. Hence I can't think this would get a lot of traction, it really would be all about the jumps.
I did think about +50% for solo jumps, and 25% for combinations where you summed up the individual jumps, but it just didn't work. E.g. for the 2A-3T/3Lz and 3Lz-3T/2A example I got max scores of (7.6 + 1.9 + 6.0 + 3.0 = 18.5) for the former, and (10.3 + 2.57 + 3.3 + 1.65 = 17.82) for the latter i.e. the max difference of 0.6 at the moment due to the different GOEs became 0.68.
Also if you went to something like +30% max GOE across the board so that the max GOEs for jumping passes became about what it is now (45 base value * 0.3 = 13.5) you would lose all the benefit of the -5 GOEs for falls and so on that the move to +/-5 would give you (the idea seems to be that really well rewarded quads say give you the same marks or even more than you get now, but you're much more heavily penalised for falls etc.). This would disappear with the 30% max, even though the 2A-3T/3Lz problem would be completely solved. A 40% max system might be a reasonable compromise if it was worked out so that well executed jumps got the same marks as now, though the penalties for falls would be smaller - I'd go for -2 marks for a fall rather than 1, that might work out OK, but it's a difficult subject, and would need someone to really go through it with a mathematical fine tooth comb.